Only T cells with the capacity to bind to self-peptide MHC complexes (self-reactivity) are included in the repertoire because of positive selection. And yet, self-reactivity is also the perilous feature of T cells that cause autoimmune disease. The distinction between useful and dangerous T cells is a matter of degree, with T cells that have a low affinity for self-MHC being useful, and those that have high affinity for self-MHC being dangerous. Yet to date, it has not been possible to study or measure the self-reactivity of diverse polyclonal populations of T cells. We propose to use a novel Nur77GFP BAC transgenic reporter mouse, in combination with tetramer enrichment and various gene deficiency models to quantitative self-reactivity in polyclonal T cells from normal and autoimmune disease prone animals. The three specific aims are: 1) to enumerate and study the T cells that are normally clonally deleted, 2) to define the contribution of CD28, medullary epithelial cells, and dendritic cells to negative selection in the thymus, and 3) to identify and characterize the self-reactive T cells that are present in diabetes-prone mice.